A Study of Incidence of AKI in Critically Ill Patients

Abstract

Background: There have been many studies to estimate the incidence of acute kidney injury (AKI) in critically ill patients. However, results were variable due to the non-usage of uniform criteria and retrospective design of most studies. There are no new studies from the developing countries looking at AKI in these patients since adoption of uniform Acute Kidney Injury Network (AKIN) criteria. Methods: In this prospective observational study from a tertiary care hospital in India, we enrolled 100 consecutively admitted critically ill patients and followed them during hospital stay. AKI was defined by AKIN criteria. Both the groups of patients, those who developed AKI and those who did not develop AKI, were then followed during the course of their hospital stay. Results: AKI occurred in 33 patients with an incidence rate of 17.3 per person year. Thirty-one out of 33 (93.9%) patients died in the AKI group, whereas 31 out of 67 (53.7%) patients died in the non-AKI group. Independent risk factors for AKI were older age (adjusted relative risk (RR) = 4.42, 95% CI = 2.57–5.23), septic shock (adjusted RR = 2.82, 95% CI = 1.43–3.80), prolonged duration of mechanical ventilation (adjusted RR = 2.35, 95% CI = 1.09–3.6), higher acute physiology and chronic health evaluation II (APACHE II) score (adjusted RR = 2.74, 95% CI = 1.28–4.13), and higher sequential organ failure assessment (SOFA) score (adjusted RR = 2.53, 95% CI = 1.04–4.08). Development of AKI was an independent risk factor for mortality (adjusted RR = 1.76, 95% CI = 1.25–1.84). Conclusion: Older patients, those with septic shock, and those requiring prolonged mechanical ventilation had increased risk for AKI. AKI was an independent predictor of mortality.

• AKIN criteria
• mortality
• developing country
• prospective
• predictors

INTRODUCTION

Acute kidney injury (AKI) is a rapid and usually reversible decline in glomerular filtration rate (GFR) that may occur in the setting of either preexisting normal kidney functions or pre-existing damaged kidney. In the ancient Greek period, the diagnosis was made only by observing a reduction in urine volume. In the modern era, however, the developments seen in fields of biochemistry and pathology have enabled us for clinicopathologic correlation and early diagnosis of AKI.¹ The Acute Kidney Injury Network (AKIN), in an effort to facilitate improved care of patients who are at risk of AKI, described AKI as “An abrupt (within 48hours) reduction in kidney function currently defined as an absolute increase in serum creatinine of more than or equal to 0.3 mg/dL (≥26.4 μmol/L), a percentage increase in serum creatinine of more than or equal to 50% (1.5-fold from baseline), or a reduction in urine output (documented oliguria of less than 0.5 mL/kg per hour for more than six hours).”²

Prior to RIFLE (risk, injury, failure, loss, and end stage renal disease) and AKIN, a variety of definitions were used to characterize AKI. This led to large variations in the reported incidence (ranging from 1–31%) and the associated mortality (19–83%)³ due to AKI. However, the differences in reported incidences of AKI and mortality could also be due to the difference in reporting of AKI among developed and developing countries, gender disparities in seeking medical attention, and seasonal variation in incidence of AKI, among many others.⁴

The proposed criteria have been evaluated in adult patients and found to be useful in characterization of AKI.⁵ We aimed to establish the incidence of AKI in critically ill patients, defined as,⁶ “all patients present with dysfunction or failure or one or more vital organs or systems,” in a tertiary care hospital in Northern India using the AKIN criteria.

METHODS

This prospective study was carried out in patients, over the age of 14 years, admitted to the medical ward and medical intensive care unit (MICU) at All India Institute of Medical Sciences, New Delhi, from August 2010 to February 2012. The inclusion and exclusion criteria are given in Table 1. Patients with serum creatinine of > 1.8 mg/dL at admission were excluded to exclude any patients with preexisting chronic kidney disease (CKD). The study was approved by the Institute Ethics Committee. Following the informed consent, information regarding the diagnosis, comorbidities, and variables during the hospital stay was noted, and the Acute Physiology and Chronic Health Evaluation III (APACHE III) score,⁷ APACHE II score,⁸ and sequential organ failure assessment (SOFA) score⁹ were calculated.

Table 1. Inclusion and exclusion criteria.

Inclusion criteria

Patients admitted to medical wards or ICU who are critically ill:

Requiring vasopressor support or

Ventilatory support or

With at least one organ failure (other than renal) or

With severe sepsis

Exclusion criteria

Hospital stay of <24 h due to any reason, including mortality

Patient taking drugs inhibiting tubular creatinine secretion: cimetidine, probenecid, and trimethoprim

Patients admitted after >48 h of admission in other hospital

Not giving consent for study

Serum creatinine of >1.8 mg/dL at admission

Serum levels of creatinine were estimated on Hitachi 717 autoanalyzer by modified Jaffe method¹⁰ at admission and every day thereafter. Urine output was measured every 6 h. Based on AKIN criteria, AKI was defined as abrupt (within 48 h) reduction in kidney function with an increase in creatinine level or a decline in urine output.² AKI was further classified as stage 1 (increase in serum creatinine by ≥0.3 mg/dL or to 1.5–1.99 times baseline), stage 2 (increase to 2–2.99 times baseline), and stage 3 (increase to ≥3 times baseline or ≥4 mg/dL with an acute rise of >0.5 mg/dL).² Sepsis was the presence of systemic inflammatory response syndrome with suspected or proven infection. Septic shock was defined as sepsis with systolic blood pressure of <90 mm Hg uncorrected by fluid infusion (Table 2).

Table 2. Definitions of sepsis, severe sepsis, and septic shock.

Sepsis	SIRS with a proven or suspected microbial etiology
Severe sepsis	Sepsis with one or more signs of organ dysfunction:
	1. Cardiovascular: arterial systolic blood pressureof ≤ 90 mm Hg or mean arterial pressureof ≤70 mm Hg that responds to administration of intravenous fluid
	2. Respiratory: PaO2/FIO2 ratio of ≤250 or, if the lung is the only dysfunctional organ, ≤200
	3. Hematologic: platelet count of < 80,000/μL or 50% decrease in platelet count from highest value recorded over the previous 3 days
	4. Unexplained metabolic acidosis: a pH <7.30 or a base deficit of >5.0 mEq/L and a plasma lactate level of >1.5 times upper limit of normal
	5. Adequate fluid resuscitation: central venous pressure of ≥8 mm Hg
Septic shock	Sepsis with hypotension (arterial blood pressure < 90 mm Hg systolic, or 40 mm Hg less than patient’s normal blood pressure) for at least 1 h despite adequate fluid resuscitation; or
	Need for vasopressor to maintain systolic blood pressure ≥ 90 mm Hg or mean arterial pressure ≥ 70 mm Hg

Notes: SIRS, Systemic inflammatory response syndrome (two or more of the following: temperature >38°C or <36°C; heart rate >90; respiratory rate >20 or pCO₂ <32; white blood cells >12,000 or <4000).

The incidence rate of AKI was calculated by dividing the number of patients developing AKI with the total number of days of follow-up till development of AKI plus total number of days of follow-up of patients in non-AKI group. Patients were evaluated for ascertaining the underlying diagnosis, progression of AKI through AKIN stages, and need for dialysis. They were followed up until discharge or death. Complete recovery of AKI was defined as normal urinalysis and normal serum creatinine. Partial recovery was defined as the presence of new onset hypertension, abnormal urinalysis (>1+ proteinuria, urine protein/creatinine ratio of >0.2 mg/mg and more than five leukocytes per high power field), or elevated serum creatinine. Patients requiring maintenance dialysis were classified as dialysis dependent.

Statistical Analysis

The incidence of AKI in critically ill patients is around 40%.^5,11,12 In order to estimate the incidence rate (hazard rate) to within 20% of true value with 95% confidence, the required sample size of critically ill patients was 100.

Results were analyzed using STATA 11 software, College station, TX, USA. Continuous data were expressed as mean (SD), median and interquartile range (IQR), and a categorical variable as number (%). The incidence rate was the number of cases per patient year. In case of continuous variables, t-test/Wilcoxon rank-sum test was applied as appropriate to see the difference in average between the groups. Chi-square test was used to see the association of categorical variables with outcome. Univariate logistic regression and stepwise multiple logistic regression were applied to calculate the unadjusted and adjusted odds ratios. Then, odds ratio was converted to relative risk (RR).¹³

RESULTS

Out of 152 patients screened, 52 were excluded including 12 patients who had serum creatinine of >1.8 mg/dL at the time of admission (Figure 1).

Figure 1. Details of patient into the study.

Incidence and Etiology

Thirty-three patients who were critically ill and had normal renal function at admission developed AKI at mean duration of 5.24 days (SD = 3.02) after admission. Thus, the incidence of AKI was 33%. The incidence rate of AKI was 17.36 per patient year. The most common etiology was considered to be septic shock (n = 22, 66.6%) (Table 3).

Table 3. Final diagnosis of patients with or without AKI.

Diagnosis	Patients with AKI (n = 33)	Patients without AKI (n = 67)
COPD	12 (36.3%)	15 (22.3%)
Hypertension	11 (33.3%)	12 (17.9%)
CAD	2 (6.0%)	9 (13.4%)
Diabetes mellitus	8 (24.2%)	10 (14.9%)
Sepsis	2 (6.0%)	10 (14.9%)
Severe sepsis	8 (24.2%)	9 (13.4%)
Septic shock	22 (66.6%)	30 (44.7%)
LRTI	20 (60.6%)	35 (52.2%)
ARDS	1 (3.0%)	8 (11.9%)
Malaria	2 (6.0%)	6 (8.9%)
Others	8 (24.2%)	10 (14.9%)

Notes: Percentages are calculated out of total number of patients in each column. AKI, acute kidney injury; COPD, chronic pulmonary airway disease; CAD, coronary artery disease; LRTI, lower respiratory tract infection; ARDS, acute respiratory distress syndrome.

Clinical Features

Table 4 shows that patients with AKI were older than those without AKI (p = 0.003). They also had significantly higher pCO₂ (p = 0.04) and low hematocrit (p = 0.05). Serum urea on day 3 was significantly higher in patients with AKI than in the non-AKI group (p = 0.0001). Similarly, on day 7, patients with AKI had significantly lower pH (p = 0.002), higher total leukocyte count (TLC) (0.0002), higher serum urea (0.0001), and lower serum bicarbonate (p = 0.03). Patients with AKI also tended to have higher APACHE III score (p = 0.003), APACHE II score (p = 0.003), and SOFA score (p = 0.001).

Table 4. Baseline characteristics of patients.

Characteristics	With AKI	Without AKI	p-Value
Age (years)	54.5 ± 20.3	42.9 ± 17.04	0.003
Gender	Male = 17	Male = 46	0.09
	Female = 16	Female = 21
Duration of illness (days)	9.4 ± 6.80	11.7 ± 13.26,	0.34
Serum creatinine (mg/dL)	0.6 ± 0.19	0.6 ± 0.16	0.31
APACHE III score	81.7 ± 21.77	68.0 ±21.39	0.003
SOFA score	9.2 ± 2.72	7.4 ± 2.57	0.001
APACHE II score	21.1 ± 6.60	16.4 ± 8.34	0.003
pH	7.3 ± 0.10	7.3 ± 0.08	0.13
pCO2	58.7 ± 29.30	46.2 ± 24.14	0.04
pO2	74.8 ± 27.51	83 ± 54.65	0.42
Serum bilirubin (mg/dL)	1.5 ± 1.25	1.06 ± 1.16	0.053
Hematocrit (%)	27.5 ± 7.27	30.5 ± 7.72	0.05
TLC (per mm^3)	18,766 ± 9129	18,976 ± 5064	0.90
Serum urea (mg/dL)	48.5 ± 14.36	44.85 ± 19.06	0.32
SGOT (IU)	61.0 ± 42.56	108.2 ± 242	0.27
SGPT (IU)	60.9 ± 42.25	71.7 ± 136.33	0.65
Sodium (meq/L)	138.9 ± 10.57	139.6 ± 9.86	0.73
Potassium (meq/L)	3.5 ± 0.758	3.6 ± 0.78	0.64
Serum albumin (g/dL)	2.9 ± 0.65	3.0 ± 0.58	0.52
Bicarbonate (meq/L)	17.9 ± 3.79	18.7 ± 4.91	0.49
Uric acid (mg/dL)	3.8 ± 2.38	3.5 ± 1.50	0.48

Notes: AKI, acute kidney injury; APACHE III, acute physiology and chronic health evaluation III; SOFA, sequential organ failure assessment; TLC, total leukocyte count.

Initially, the distribution of patients in AKI stages were 21 (63.6%) in stage 1, 9 (27.2%) in stage 2, and 3 (9%) in stage 3. Maximum AKI stages of the patients were stage 1 [3 (9%)], stage 2 [4 (12.1%)], and stage 3 [26 (78.7%)]. Dialysis was required in eight patients (24.2%)—two patients received hemodialysis and six were initiated on acute peritoneal dialysis. The mean duration of ICU stay was 7.7 days for patients with AKI (SD ± 4.21) and 5.9 days for patients without AKI (SD ± 2.64). In the AKI group, 32 (96.9%) patients were mechanically ventilated, whereas in the non-AKI group, 64 (95.5%) patients were mechanically ventilated. In the AKI group, 81.8% of patients required vasopressor support, whereas in the non-AKI group, 53.7% of patients required vasopressor support (p = 0.006). Patients with AKI also had higher number of days of urinary catheterization than those without AKI (p = 0.011). No difference was seen among both the groups in terms of nephrotoxic drug use (p = 0.330).

Outcome

Out of 100 patients, 67 died during hospital stay. The mortality was 93.9% in the AKI group and 53.7% in the non-AKI group. Both the patients who survived in the AKI group attained complete renal recovery. On logistic regression analysis, the independent risk factors for AKI included age of >45 years, septic shock, mechanical ventilation for >4 days, APACHE II score of >16, and SOFA score of >7.4 (p < 0.05) (Table 5). On univariate analysis, the risk factors for mortality in critically ill patients were age of >45 years (p = 0.05), female sex (p = 0.001), lower Glasgow Coma Scale (GCS) at admission (p = 0.004), low platelet count (p = 0.021), higher APACHE III score (p = 0.007), prolonged length of hospital stay (p = 0.0001), low PaO₂/FiO₂ ratio (p = 0.05), development of AKI (p = 0.001), and prolonged duration of mechanical ventilation (p = 0.003). On multivariate analysis, female sex, duration of mechanical ventilation for >3 days, APACHE III score of >64, and development of AKI were the only independent predictors of mortality (Table 6).

Table 5. Risk factors for AKI.

Variable	With AKI (n = 33) f (%)	Without AKI (n = 67) f (%)	p-Value	Univariate RR (95% CI)	Adjusted RR (95% CI)
Age (years)
<45	8 (18.1)	36 (81.8)	0.005	1	1
>45	25 (44.6)	31 (55.3)		2.45 (1.32–3.69)	4.42 (2.57–5.23)
Septic shock
No	11 (22.9)	37 (77)	0.039	1	1
Yes	22 (42.3)	30 (57.6)		1.84 (1.02–2.77)	2.82 (1.43–3.80)
MV (days)
<4	9 (21.4)	33 (78.5)	0.032	1	1
>4	23 (42.5)	31 (57.4)		1.98 (1.06–3.02)	2.35 (1.09–3.60)
APACHE II
<16	11 (22)	39 (78)	0.021	1	1
>16	22 (44)	28 (56)		2.07 (1.12–3.2)	2.74 (1.28–4.13)
SOFA
<7.4	8 (19.1)	34 (80.9)	0.014	1	1
>7.4	25 (43.1)	33 (56.8)		2.25 (1.2–3.45)	2.53 (1.04–4.08)

Notes: AKI, acute kidney injury; RR, relative risk; MV, mechanical ventilation; APACHE II, acute physiology and chronic health evaluation II; SOFA, sequential organ failure assessment. n = total number of patients and f = frequency of distribution.

Table 6. Risk factors for mortality in critically ill patients.

Variables	Discharge (n = 33) f (%)	Death (n = 67) f (%)	p-Value	Univariate RR (95% CI)	Adjusted RR (95% CI)
Sex
Male	28 (44.4)	35 (55.5)	0.001	1	1
Female	5 (13.5)	32 (86.4)		1.55 (1.23–1.70)	1.55(1.07–1.73)
Duration of MV (days)
<3	18 (42.8)	24 (57.1)	0.001	1	1
>3	11 (20.3)	43 (79.6)		0.36 (0.12–0.71)	1.60 (1.24–1.71)
Developed AKI
No	31 (46.2)	36 (53.7)	0.001	1	1
Yes	2 (6.0)	31 (93.9)		1.74 (1.44–1.83)	1.76 (1.25–1.84)
APACHE III score
<64	21 (50)	21 (50)	0.002	1	1
>64	12 (20.6)	46 (79.3)		1.58 (1.22–1.80)	1.64 (1.11–1.89)

Notes: RR, relative risk; MV, mechanical ventilation; AKI, acute kidney injury; APACHE III, acute physiology and chronic health evaluation III. n = total number of patients and f = frequency of distribution.

DISCUSSION

In this first study from India looking prospectively at hospitalized critically ill patients using AKIN criteria, we found that the incidence of AKI in critically ill patients was 33% with an incidence density of 17.36 cases per person year of follow-up. While most patients developed stage 1 disease at diagnosis, there was progressive renal dysfunction with 78.7% patients reaching stage 3 disease and 24.2% patients required renal replacement therapy. Older patients, those with septic shock, and those requiring prolonged mechanical ventilation were at increased risk of AKI. The presence of AKI resulted in prolonged ICU stay and was associated with 1.8 times increase in mortality. All the patients who received dialysis died and only two patients who did not receive dialysis survived.

In 2007, the AKIN proposed to replace the term acute renal failure with AKI to encompass the dynamic nature of the entity and proposed criteria for the same.² Prior to the AKIN and RIFLE criteria, significant heterogeneity existed among the reported incidence of AKI. The BEST kidney group¹⁴ reported a period prevalence of AKI to be 5.7% with the requirement of RRT in approximately two-thirds of patients who developed AKI (4% of total), whereas in the PICARD study,¹⁵ RRT was required in 64% patients. The proposed classifications, the RIFLE¹⁶ and AKIN² criteria, have been validated as diagnostic and prognostic tools in critically ill adult patients with AKI.^5,17 Using the AKIN criteria, Joannidis et al.⁵ found the incidence of AKI in critically ill patients to be 28.5%. Use of the same criteria yielded an incidence of 24.4% in a study from Hungary.¹¹ Ostermann et al., in a retrospective study of 22,303 patients admitted in adult ICU, found that 7898 (35.4%) had AKI.¹² Using same criteria, we found that 33% of critically ill patients developed AKI during hospital stay.

The etiologies of AKI in critically ill patients are studied in various studies. The most common cause quoted in the literature is septic shock.^14,18,19 Other factors that are associated with increased risk in the development of AKI are older age, preexisting CKD, nephrotoxic drugs, and failure of organ other than kidney.²⁰ In patients with sepsis, the variables associated with increased risk of developing AKI are older age, higher serum creatinine at admission (although under normal range), higher central venous pressure (CVP) on day 1, presence of liver failure (serum bilirubin > 1.5), and pH <7.3 on day 1.^21,22 We identified the following variables as predictors of AKI in critically ill patients in multivariate analysis: older age, presence of septic shock, higher duration of mechanical ventilation, higher APACHE II score, and higher SOFA score.

The occurrence of AKI has significant implications in terms of short- and long-term morbidity and mortality. The overall in hospital mortality rates of patients with AKI ranges from 27% to 72%.^11,23–26 The mortality rate of 93.9% observed in this study reflects a sicker cohort, where 96% of all patients were mechanically ventilated. In addition, 81.8% of patients in the AKI group required vasopressor support, whereas 53.7% of patients in the non-AKI group required vasopressor support (p = 0.006). Also, the patients in the AKI group had higher mean age, APACHE score, and SOFA scores at baseline as compared to the non-AKI group. The total mortality rate in this cohort was 67%, which is comparable to that reported in other studies, which report a similar rate.^27,28 This study also shows that development of AKI during hospital stay is an independent predictor of mortality (adjusted RR = 1.76, 95% CI = 1.25–1.84).

Along with development of AKI, other factors that were significant in multivariate analysis as predictors of death are female sex, longer duration of mechanical ventilation, and higher APACHE III score at admission. The higher mortality in female sex seen in this study may be due to smaller sample size of the study. This study was underpowered to detect a difference in mortality between the three AKIN stages of AKI.

This study has a few limitations. This is a single-center study in a group of patients who were sicker and had much comorbidity. Further, exclusion of patients with serum creatinine >1.8 mg/dL may have resulted in erroneous exclusion of patients with unknown but early chronic disease.

This prospective study provides data on the incidence of AKI in hospitalized critically ill patients. It emphasizes that the incidence of AKI is higher in critically ill patients, especially the elderly with septic shock and prolonged duration of mechanical ventilation. The presence of AKI results in increased mortality. Further prospective studies are required to examine long-term impact of AKI on renal outcome.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.